quantization / build.toml

Commit History

Fix undefined symbol on CUDA 11.8

3313895

danieldk HF Staff commited on about 1 month ago

Set CUDA minimum versions for hopper/blackwell cutlass kernels

3370704

danieldk HF Staff commited on Jul 3

Temporarily disable ROCm

d5fd23d

danieldk HF Staff commited on Jul 4

Sync to vLLM 20250627

8aa00a3

danieldk HF Staff commited on Jul 2

Sync capabilities with upstream

59b2fef

danieldk HF Staff commited on Apr 16

Add support for ROCm

3c8bb73

danieldk HF Staff commited on Mar 31

Update for build.toml changes

116e562

danieldk HF Staff commited on Feb 26

Sync with vLLM

0da5bf5

danieldk HF Staff commited on Jan 16

Add full Marlin support and tests for Marlin/CUTLASS

165b25c

danieldk HF Staff commited on Dec 11, 2024

Add GPTQ-Marlin

c31b5ce

danieldk HF Staff commited on Dec 10, 2024

Add `scaled_(int|fp8)_quant` and `fp8_marlin_gemm`

5c6fb68

danieldk HF Staff commited on Dec 9, 2024

Add cutlass_w8a8

b4cad21

danieldk HF Staff commited on Dec 9, 2024

Commit History

Fix undefined symbol on CUDA 11.8 3313895

Set CUDA minimum versions for hopper/blackwell cutlass kernels 3370704

Temporarily disable ROCm d5fd23d

Sync to vLLM 20250627 8aa00a3

Sync capabilities with upstream 59b2fef

Add support for ROCm 3c8bb73

Update for build.toml changes 116e562

Sync with vLLM 0da5bf5

Add full Marlin support and tests for Marlin/CUTLASS 165b25c

Add GPTQ-Marlin c31b5ce

Add `scaled_(int|fp8)_quant` and `fp8_marlin_gemm` 5c6fb68

Add cutlass_w8a8 b4cad21